Comprehensive analysis of formin gene family highlights candidate genes related to pollen cytoskeleton and male fertility in wheat (Triticum aestivum L.)

Formin, a highly conserved multi-domain protein, interacts with microfilaments and microtubules. Although specifically expressed formin genes in anthers are potentially significant in research on male sterility and hybrid wheat breeding, similar reports in wheat, especially in thermo-sensitive genic male sterile (TGMS) wheat, remain elusive. Herein, we systematically characterized the formin genes in TGMS wheat line BS366 named TaFormins (TaFHs) and predicted their functions in inducing stress response. In total, 25 TaFH genes were uncovered, majorly localized in 2A, 2B, and 2D chromosomes. According to the neighbor-joining (NJ) method, all TaFH proteins from wheat and other plants clustered in 6 sub-groups (A-F). The modeled 3D structures of TaFH1-A/B, TaFH2-A/B, TaFH3-A/B and TaFH3-B/D were validated. And different numbers of stress and hormone-responsive regulatory elements in their 1500 base pair promoter regions were contained in the TaFH genes copies. TaFHs had specific temporal and spatial expression characteristics, whereby TaFH1, TaFH4, and TaFH5 were expressed highly in the stamen of BS366. Besides, the accumulation of TaFHs was remarkably lower in a low-temperature sterile condition (Nanyang) than fertile condition (Beijing), particularly at the early stamen development stage. The pollen cytoskeleton of BS366 was abnormal in the three stages under sterile and fertile environments. Furthermore, under different stress levels, TaFHs expression could be induced by drought, salt, abscisic acid (ABA), salicylic acid (SA), methyl jasmonate (MeJA), indole-3-acetic acid (IAA), polyethylene glycol (PEG), and low temperature. Some miRNAs, including miR167, miR1120, and miR172, interacts with TaFH genes; thus, we constructed an interaction network between microRNAs, TaFHs, phytohormone responses, and distribution of cytoskeleton to reveal the regulatory association between upstream genes of TaFH family members and sterile. Collectively, this comprehensive analysis provides novel insights into TaFHs and miRNA resources for wheat breeding. These findings are, therefore, valuable in understanding the mechanism of TGMS fertility conversion in wheat.

Conclusions: Collectively, this comprehensive analysis provides novel insights into TaFHs and miRNA resources for wheat breeding. These findings are, therefore, valuable in understanding the mechanism of TGMS fertility conversion in wheat.
Keywords: Wheat, Thermo-sensitive genic male sterile, Formin gene family, Cytoskeleton, MiRNA Background Eukaryotic cells possess a dynamic actin cytoskeleton that controls cell growth, division, and polarity formation by alternating its global and filamentous state in response to developmental and environmental stimuli [1]. Formins (formin homology proteins), a group of proteins involved in actin polymerization, are associated with the fast-growing end (bared end) of actin filaments, potentially mediating a series of cellular functions, including polarity, division, cytokinesis, migration, among others. Several formins have been identified in plants. The formin proteins are characterized by two formin homology (FH) domains, FH1 and FH2. The FH1 (formin homology 1) domain harbors a distinct polyproline-rich region that binds to a crucial actin-binding proteinprofilin or actin/profilin complex. Profilin is a vital monomer actin-binding protein and can interact with formin via the FH1 domain. The FH2 (formin homology 2) domain is characterized by a key sequence that interacts with actin [2]. Plant formins occur in two categories, type I and type II, based on the sequence homology of their FH (formin homology) domains. Type I formins have a transmembrane (TM) domain at their N-terminus followed by the C-terminal FH1 and FH2 domains, whereas type II formins do not have an N-terminal TM domain but carry an N-terminal phosphatase and tensinrelated (PTEN)-like domain besides the conserved FH1 and FH2 domains [3].
Also, formins have multiple members in each species; for example, 22 formin family members are present in the Arabidopsis genome. Specifically, AtFH1, AtFH5, AtFH6, and AtFH8 have been described in Arabidopsis and their functions explored in vivo. For instance, AtFH1 overexpression in pollen tubes induced the formation of arrays of actin cables, projecting into the cytoplasm from the cell membrane. Interestingly, AtFH1 expression induced tube broadening, growth depolarization, and growth arrest in transformed pollen tubes. With these observations, it was suggested that AtFH1 regulation of actin polymerization could be necessary for the polarized pollen cell growth process [4]. The AtFH5-GFP fusion protein accumulated in the cell plate, essential for cell division. Furthermore, AtFH6 regulated polarized growth by tuning the assembly of actin cables [5]. Other reports revealed that AtFH8 affects root and root hair development by alternating actin cytoskeleton distribution [6,7]. Collectively, formins function as key factors potentially bridging several vital pathways related to plant pollen growth and development, particularly cytoskeleton distribution. Notably, formins have been adequately characterized in animals, yeast, and Arabidopsis. For example, the first gene encoding forming-the mouse limb deformity (ld) gene in 1982, the mutation of which led the mice to fail to"form" proper limbs and kidneys, and a large family of formin homology proteins were related with the field of actin cytoskeleton [8,9]. However, the formin gene family members are yet to be uncovered in the wheat genome. Also, it is not clear whether formin genes are related to pollen development, especially in the distribution of microspore cytoskeleton in wheat.
Wheat is one of the most important staple food crops in China and globally in terms of production and consumption [10]. Given that wheat production varies with climate and other environmental conditions, application of molecular biology and genetics methods is critical for advancing the stress tolerance and quality of wheat. Thermo-sensitive genic male sterile wheat (TGMS) lines, such as BS366, are of particular significance in a twoline hybrid system. The lines are highly efficient in breeding [11] because their fertility is strictly temperature-dependent [12][13][14]. Hybrid seeds can be produced under sterile conditions (TGMS line as the maternal plant), whereas the TGMS line can be reproduced under fertile conditions [15]. Thus, the BS366 is an ideal material for studying the genetic mechanism of TGMS and realize the effective adoption of hybrid wheat. In our previous cytological findings, the male sterility in the TGMS wheat cultivar BS366 was primarily induced by the disordered cytoskeleton distribution during meiosis of microsporocytes [16]. However, no study has reported on the contribution of formin gene family members related to pollen cytoskeleton distribution in TGMS wheat. Additionally, miRNAs are one type of small RNAs, which directly mediate the expression of target genes through post-transcriptional negative regulation [17]. Further information on miRNA association with the formin gene family is scanty. Therefore, there is a need to comprehensively explore the formin gene family in wheat and elucidate the interaction between TaFHs and miRNA; this would reveal their potential roles during anther development inBS366.
In this study, after identifying the formin genes of wheat (TaFHs), we evaluated their characteristics for the physicochemical, phylogenetic relationship, 3D structure, motif structure, exon/intro structure, chromosomal location, miRNA regulation, and expression patterns in different tissues and different development stages of anthers in the TGMS wheat line, BS366. The results presented will enrich our knowledge of the TaFH genes family and offer the theoretical basis and novel candidate genes for improving and developing the male sterility wheat lines. It is imperative for us to understand the fertility pathway of TGMS.

Identifying formin genes in wheat
Two strategies were adopted to mine for formin genes in T. aestivum L. BLASTP using the formin of A. thaliana proteins as queries and HMMER searches using the formin domain as a query against the T. aestivum L. protein database. NCBI-CDD and SMART were used to confirm whether the formin domains were present. In total, 25 formin genes were identified, denoted as TaFH1 to TaFH10, based on their location on the chromosomes. Based on the physicochemical analysis of the protein sequence encoded by the wheat formin gene, there were significant differences in physical and chemical properties across the members of the wheat formin family. The protein comprised 306-1403 amino acids in length, with a molecular weight of 24.40-153.70 and 5.3-9.86 isoelectric points. Subcellular localization predictions revealed that TaFH proteins functioned in cytoplasmic, nuclear, vacuolation, plasma membrane, chloroplast, and endoplasmic reticulum (Table 1). Different characteristics of TaFH genes and proteins were also revealed. Results demonstrated that different TaFH proteins exerted potentially distinct biological functions.

Phylogenetic tree and domain analysis of TaFHs
To elucidate the phylogeny of formin proteins, we adopted the MEGA software to analyze the 25 identified formin genes using an evolutionary tree. Based on the information on the phylogenetic relationship, 25 TaFH genes were clustered into 10 groups (Fig. S1). Results demonstrated that 25 TaFH genes could be classified into four groups; the large group could then be subdivided into four groups. The conservative domains of these domain were the regions where genes function. The TaFHs structure were obtained and the number of exons and introns were highly divergent (Fig. S2). Most members of TaFHs shared the similar patterns of exon/ intro structure, including intro phase, intro number and exon length.
To reveal the functional information of TaFH genes, we constructed a phylogenetic tree based on the comparison among the T. aestivum, P. patens, S. moellendorffii, P. trichocarpa, S. bicolor, B. distachyon, A. thaliana, and O. sativa (Fig. 1). All the formin proteins clustered into six groups, except for the PpFH2, PpFH4 and PpFH5. In total, 25 sequences of 10 members of the wheat formin gene family clustered into six subgroups. Moreover, we found that the TaFH proteins were clustered into the same clades with some BdFH, SbFH and OsFH proteins. For example, TaFH2, BdFH12, SbFH12 and OsFH8 were clustered into group A, BdFH1, SbFH6 and OsFH5 were clustered with TaFH3 on the group B, TaFH4 and TaFH5 were clustered into the group D with BdFH2, BdFH11, SbFH5, SbFH11, OsFH4 and OsFH3, and TaFH9 were clustered into the group E with BdFH7, SbFH10 and OsFH1. These showed that TaFH proteins had a close correlation with those in B. distachyon, S. bicolor and O. sativa and shared a low similarity with those in P. patens.
In this study, eight TaFHs representative proteins were modeled using homology to predict the 3D structure. The Phyre 2 server and PyMDLWin been used to deduced and visualized, which were shown in Fig. S3. The results also shown that protein structure of A, B and D homoeologous copies with TaFH1-A/B, TaFH2-A/B, TaFH3-A/B and TaFH9-B/D had shown significant similarity. The 3D structure of TaFH proteins laid a foundation to explor the complex mechanism of formins at the structure level.

Chromosomal assignment of TaFH genes
Map files were constructed based on the location of the wheat formin gene family on the chromosome. A chromosome distribution map was drawn using the MapInspect software. A total of 25 wheat formin genes were located on the chromosomes (Fig. 2). Three copies of TaFH1, TaFH2, TaFH3, TaFH4, TaFH6, and TaFH10 were located on sub-genomes A, B, and D of wheat chromosomes 1, 2, 3, and 6. Two copies of TaFH5, TaFH7, and TaFH9 were located on sub-genomes of wheat chromosomes 2, 4, and 6, respectively. TaFH8 had one copy on the short arm of chromosome 5A. Four formin genes were found on chromosomes 2A and 2B, and three formin genes on chromosome 2C, existing in clusters. However, no formin gene was present on 7A, 4B, 5B, 7B, 5D, and 7D, and this demonstrated a widespread but uneven distribution of the wheat formin genes on the wheat chromosome.

MicroRNA targeting prediction of TaFH genes
To further understand the function of TaFH genes, we evaluated the binding association of the TaFH genes with miRNAs. Results revealed that 19 of the 25 TaFH genes harbored targets for 13 different miRNAs ( Fig. 3 and Table S1). We also found that tae-miR167a, tae-miR167b, and tae-miR167c-5p potentially target TaFH2-A, TaFH2-B, and TaFH2-D. Indeed, Tae-miR167 has been shown to target other wheat genes. For example, tae-miR167a directly targeted Auxin Response Factor 8 (TaARF8), which possibly regulated auxin biosynthesis by impeding the downstream genes. Also, tae-miR167a overexpression in Arabidopsis lowered the expression levels of its targets (AtARF6 and AtARF8) and induced male sterility phenotypes [18]. Our results suggested that tae-miR167b might target TaFH1-A, TaFH1-B, and TaFH1-D. Tae-miR1120b-3p has been shown to target TaFH3-A and TaFH3-B, and possibly LRR, thereby regulating the jasmonate biosynthesis or auxin response pathway associated with pollen development. A previous study found that the Q gene, involved in bread wheat spike architecture, was suppressed by tae-miR172 [19]. In a separate study, Tae-miR444a targeted TaFH9-D and regulated floral patterning and development via the transcription factor MADS-box [20]. Collectively, the findings of this study could help to further understand the regulation between miRNA and formin in the related TGMS pathway.

Promoter analysis of TaFH genes
To establish how the expression levels of TaFH genes responded to stress stimuli, 1.5-kb upstream promoter regions of TaFH genes were scanned for stress-related cis-regulatory elements using the Plant CARE online service. Eight hormone-responsive regulatory elements (ABRE, TGA-element, AuxRR-core, TGACG-motif, CGTCA-motif, P-box, GARE-motif, and TCA-element) associated with ABA, auxin (IAA), methyl jasmonate (MeJA), gibberellin (GA), and salicylic acid (SA) responses were identified in the promoter region of TaFHs ( Fig. 4) Additionally, five stress-responsive regulatory elements (TC-rich repeats, MBS, ARE, GC-motif, and LTR) associated with defense/stress and low-temperature responses were identified in the TaFH promoter regions. Different types and numbers of regulatory elements were present in the distinct TaFH promoters. This implied that TaFH genes are potentially associated with the response to various stress and hormone treatments by mediating different regulatory mechanisms.
Tissue/organ-specific expression profiles of TaFH genes We explored the expression of TaFH genes in different tissues of TGMS wheat line BS366, including the root, stem, leaf, pistil, stamen, and glume. All the TaFH genes were constitutively expressed in all the tissues. The expression of TaFH1 in pistils was the highest, whereas that in stamens was about 45 folds more than in the roots (Fig. 5). The expression of TaFH6, TaFH3, and TaFH8 in pistils was markedly higher than those in other tissues, about 20, 3 and 1.2 folds more than in roots. The expression of TaFH4 and TaFH5 in stamens was the highest, about 30 and 120 folds more than that in roots, implying that TaFH1, TaFH4 and TaFH5, were potentially linked to anther development. Also, TaFH2 expression in leaves was the highest, nearly 9 folds higher than in the roots, but relatively lower in the corresponding tissues of flower organs. This suggested that TaFH2 could be mainly related to the development of leaves. The expression of TaFH7 in glume was the highest, nearly 16 folds higher than that in the roots. The expression levels of TaFH9 and TaFH10 in glumes were higher than the root, stem, leaf, pistil and stamen, about 9 and 20 folds than in roots, respectively. Thus, we speculated that TaFH7, TaFH9, and TaFH10 could be related to the opening and closing of glumes. Further analysis revealed that 10 members of the wheat formin gene family were mainly associated with the development of the above-ground tissues of BS366, mostly the floral organs. In particular, they contributed to the development of anthers and glumes, and other metabolic processes in TGMS wheat.

Analysis of TaFH genes family expression patterns in fertility sensitive stages under different fertility conditions in BS366
Based on previous reports, the fertility sensitive stage of BS366 was the meiosis stage of microspore mother cells [10,16]. Herein, we selected three key meiotic stages (1, 2, 3) for expression pattern analysis. The role of the wheat formin gene family in pollen fertility was further elucidated by analyzing the expression of TaFHs in the three stages of stamen development in BS366 under Fig. 4 Analysis of specific cis-elements in promoters of TaFHs. The 1.5 k-bp promoter sequences of corresponding TaFH genes were used to analyze specific hormone-related cis-elements and stress-responsive regulatory elements, which are color-coded. ABRE: abscisic acid responsive element; ARE: antioxidant response element mediating transcriptional activation of genes exposed to oxidative stress; AuxRR-core: auxinresponsive element; CGTCA-motif and TGACG-motif: MeJA-responsive element; GARE-motif and P-box: gibberellin-responsive element; GC-motif: enhancer-like element involved in anoxic specific inducibility; LTR: low-temperature responsive element; MBS: drought-inducible element; TCAelement: salicylic acid responsive element; TGA-element: auxin responsive element; TC-rich repeat: defense and stress responsive element different fertility environments (fertile and sterile conditions). The expression levels of TaFH2 and TaFH4 in the corresponding three stages of stamen development under sterile conditions were lower than those under fertile conditions (Fig. 6). TaFH1 expression was higher at the early stage (stage 1) under sterile conditions. The expression of TaFH1, TaFH3, TaFH5, and TaFH10 were significantly higher at stage1, but stage 2 under sterile conditions. On the other hand, TaFH6 expression was higher at the early stages of stamen development (stage 1 and 2) under sterile conditions. TaFH7 and TaFH8 showed no obvious expression changes at the three stages under both fertility and sterile conditions. However, TaFH9 expression was lower under sterility conditions than in fertility conditions at stage 1 and 2. These results implied a potential association of TaFH genes with stamen development and fertility transformation via particular pathways.

Distribution of cytoskeleton in meiosis of microsporocytes
Compelling evidence shows that the pollen cell cytoskeleton of BS366 is abnormal at low temperature, and the formin gene is highly critical in the development of the pollen cell cytoskeleton [15,16]. In the present study, to establish any change in the cytoskeleton of the TGMS 366 in the fertility and sterility environment, the pollen cell was evaluated in three key meiotic stages. We found abnormal cytoskeleton distribution in male-sterile wheat pollen cells (Fig. 7). The pollen cell shrank, and the cytoskeleton was loose and disarranged in stage 1. The cell plate was absent, and the microtubule and microfilaments nonpolar terminus were vague in stage 2. In stage 3, the microtubule and microfilaments disappeared. Collectively, these findings suggested that differential expression of TaFH genes might cause the abnormal pollen cytoskeleton of BS366 under low temperatures.

Expression profiles of TaFH genes under different exogenous phytohormone treatments and abiotic stress
Abscisic acid (ABA), gibberellin (GA), indole-3-acetic acid (IAA), methyl jasmonate (MeJA), and salicylic acid (SA) are the main plant growth regulators that exert critical functions in plant growth and metabolism. By exploring the cis-element of TaFH, we found that TaFH had different types and numbers of regulatory elements.
To affirm the function of TaFH, we conducted a qRT-PCR assay to examine the expression pattern of the formin genes in the leaves of BS366 following different hormone treatments (Fig. 8). According to the results, all the formin genes slightly responded to the ABA, GA, IAA, MeJA, and SA treatments. Specifically, TaFH3, TaFH4, TaFH5, TaFH7, TaFH8, TaFH9, and TaFH10 were no obvious expressed changing under abscisic acid-induced stress, the expression of TaFH1 and TaFH6 was initially inhibited but later rose and peaked at 8 h, after which it began to show a wave-like decreasing trend and then rose again. Under auxin (IAA) stress, TaFH1, TaFH2, TaFH3, TaFH4, TaFH5, TaFH6, TaFH7, TaFH8, and TaFH10 were initially overexpressed, but their expression later decreased. The expression of TaFH1, TaFH2, and TaFH8 peaked at 4 h, whereas TaFH9 expression was impeded and then increased, and later peaked at 12 h. The expression of TaFH2, TaFH3, TaFH4, TaFH5, TaFH6, TaFH7, TaFH8, TaFH9, and TaFH10 demonstrated a certain upward trend under gibberellin stress. The expressions of TaFH2 and TaFH10 rose significantly at 2 h after gibberellin stress, then decreased and increased again. TaFH3, TaFH4, TaFH5, TaFH6, TaFH7, TaFH8, and TaFH9 showed a wave-like increasing trend, then decreased and rose again. The expression of TaFH4, TaFH5, TaFH6, TaFH7, and TaFH8 peaked at 24 h, whereas that of TaFH9 peaked at 4 h. Under methyl jasmonate stress, TaFH1 expression was inhibited but later increased at 12-24 h. The expression of TaFH2, TaFH3, TaFH4, and TaFH5 was inhibited within 24 h, and that of TaFH6 and TaFH10 showed an upward and downward trend, then peaked at 24 h. Further, the expression of TaFH7 was inhibited at 2 h, then increased and later decreased. TaFH8 expression peaked at 8 h, increased at 2 h, but then decreased, increased, and then decreased again at 12 h; the relative expression peaked at 2 h. The expression of TaFH9 was enhanced, then inhibited, and peaked at 8 h. Under salicylic acid stress, the relative expression of TaFH2, TaFH5, TaFH7, TaFH8, TaFH9, and TaFH10 was higher at 2 h compared to 0 h. TaFH2 expression peaked at 2 h, while TaFH5 peaked at 8 h. The expression of TaFH7, TaFH8, TaFH9 and TaFH10 all peaked at 4 h. Besides, the relative expression of TaFH1, TaFH3 and TaFH6 decreased at 2 h, whereas that of TaFH1 decreased at 8 h. Also, the expression of TaFH3, TaFH4, and TaFH6 peaked at 4 h. These results demonstrated that formin genes might be central to hormone regulation in the BS366 wheat.
To evaluate the response mechanism of the formin gene family to cold stress, drought, and high salt, we employed qRT-PCR to analyze the expression pattern of Initially, the expression of TaFH2, TaFH3, TaFH4, TaFH5, TaFH8, TaFH9, and TaFH10 was inhibited under cold stimulation, but later rose. The expression of most TaFH genes peaked at 12 h; however, the expression of TaFH1, TaFH6, and TaFH7 rose and then decreased post-cold stimulation. High salt stress impeded the expression of TaFH1, TaFH4, TaFH7, TaFH8, TaFH9, and TaFH10, whereas the expression of TaFH2, TaFH3, TaFH5, and TaFH6 showed an opposite trend. Under abiotic drought stress, except for TaFH7 and TaFH8, the expression of TaFH2, TaFH3, and TaFH9 peaked at 8 h and then decreased. These results indicate that TaFH genes might respond to different plant hormones and participate in a complex regulatory network in thermo-sensitive male sterile line BS366.

Discussion
The relationship between formin genes and cytoskeleton in wheat Formins initiate actin assembly and control key cytoskeleton genes. They regulate the nucleation of actin and mediate profilin and Arp2/3 during cell division. Also, they participate in vegetative and reproductive plant growth [21]. Previously, [15] found abnormal cytoskeleton formation in a TGMS line under sterile environment. In this study, we examined the cytoskeleton of the pollen-sensitive stages in BS366 using scanning electron microscopy and found aberrant nuclear division in the pollen cells. Real-time quantification analysis of the formin genes revealed different expression of TaFH1, TaFH2, TaFH3, TaFH4, TaFH5, TaFH6, TaFH9 and TaFH10 under sterile and fertile conditions at these stages. Therefore, we concluded that the formin genes TaFH1, TaFH2, TaFH3, TaFH4, TaFH5, TaFH6, TaFH9, and TaFH10 might affect the cytoskeleton, resulting in abnormalities during meiosis and thus male sterility in TGMS BS366.

Effect of microRNA regulation via formin genes on the fertility of TGMS line
MiRNAs are endogenous, small, non-coding RNAs, exerting critical roles in the regulation of hormone signaling pathways and cellular functions during developmental processes [22]. Of note, miRNAs participate in the regulation of male fertility in crops [23]. Particularly, miR167 plays a dominant role in plant development. In Arabidopsis, miR167 directly targets ARF6 and ARF8, which regulates jasmonate biosynthesis by impeding downstream genes, thereby affecting pollen development. In rice, miR167 potentially targets ARF6, 8, 12, 17, and 25 genes. Notably, ARF8 can interact with OsGH3-2. Besides, OsmiR167 responds to exogenous auxin via its target gene OsARF8. Abnormal OsGH3-2 expression induces morphological aberrations, including short plants and small panicles [24][25][26]. Studies have shown that TaemiR167a, TaemiR167b, and TaemiR167d are abnormally expressed in the sensitive stage of pollen in wheat line BS366 [27,28]. TaemiR167a targeted AtARF6 and AtARF8 upon its translation into Arabidopsis, and influenced the synthesis of IAA and JA, thereby regulating the development of anthers [18].
In the present study, by analysis the relationship between the miRNA and formins, we found that taemiR167a targeted TaFH2-A, TaFH2-B, and TaFH2-D; taemiR167b targeted TaFH1-A, TaFH1-B, TaFH1-D, TaFH2-A, TaFH2-B, and TaFH2-D; whereas taemiR167c-5p targeted TaFH2-A, TaFH2-B, and TaFH2-D. As such, we suggested that taemiR167 might regulate the formin genes in pollensensitive stages and influence BS366 fertility; leading to abnormal anther dehiscence. Moreover, the gene FBA (fructose-1, 6-bisphosphate aldolase) was predicted to be the target of miR1120. Of note, miR1120 could be regulated by ABA under low-temperature treatment, which affected the expression of FBA, then regulated the sugar metabolism to mediate ATP generation [28]. In a previous report, the OsFH5 mutants were characterized by aberrant inflorescence. The cytological analysis demonstrated that OsFH5 mutants exhibited severe cell elongation defects and abnormal microtubule arrays [29]. Herein, miR1120 was predicted to target TaFH3 (Fig. 3), OsFH5, and TaFH3, clustered into group B (Fig. 1). Also, the cytoskeleton was abnormal during anther development. Thus, we speculated that miR1120 potentially regulated the energy synthesis and then impacted the pollen cell division and the fertility of the TGMS line BS366 under low-temperature treatment.

Expression and functions of formin genes in response to abiotic stressors
A previous study by [27] found abnormal levels of IAA during TGMS366 pollen development. Other reports have also revealed that MeJA, JA, and SA influence the pollen spill out, which results in the sterility of BS366 [30,31]. In this study, we isolated the formin gene family, including 25 members from wheat ( Table 1). The formin genes exhibited different expression levels in different BS366 tissues. For instance, the relative expression of TaFH4 and TaFH5 in stamen was higher compared to other tissues. Also, the relative expression of TaFH1, TaFH3, TaFH6, and TaFH8 in pistil was higher than in other tissues, suggesting that TaFH1, TaFH3, TaFH4, TaFH5, TaFH6, and TaFH8 potentially contribute to the sterility of BS366. Furthermore, the relative expression of TaFH1, TaFH2, TaFH3, TaFH4, TaFH5, TaFH6, TaFH9, and TaFH10 was different in the sensitive stage between the fertility and sterility conditions, implying that they might play a regulatory role in wheat fertility. Cis-regulatory sequences, such as enhancers and promoters, influence development and physiology by regulating gene expression [32]. Cis-regulatory elements, functioning as important molecular switches, are associated with the regulation of gene transcription under external stimuli [33]. To comprehensively elucidate the functions of TaFH genes, we assessed cis-acting regulatory elements and expression profiles under various stresses. Eight responsive regulatory elements and five stress-responsive regulatory elements were identified in the TaFH promoter region. These findings demonstrated that the TaFH genes respond to stress and hormone treatment variations via different mechanisms. The TaFHs contained different types and numbers of cis-acting regulatory elements in each promoter region. These genes, therefore, were characterized by different regulatory functions that responded to variations in stress and hormone treatments. We also evaluated the expression profiles of TaFH genes in TGMS wheat line BS366 seedlings under different stress conditions and hormone treatments. The expression profile analysis revealed that TaFH8 and TaFH10 had similarly structured proteins upregulated by the auxin response treatment. Similar expression profiles were shown in the TaFH2 following low-temperature treatment.
Moreover, the response of formin genes to the ABA, NaCl, and PEG implied that TaFHs might exert multiple functions in plants, including roles in abiotic stress resistance and pollen fertility. Although formin is a key regulatory gene in the cell cytoskeleton, exploring the mechanisms by which it responds to low temperature, affects the cytoskeleton, and defines pollen development of wheat would be vital to reveal the actual function of TaFH.

Possible miRNA-TaFH interaction network in wheat
In this study, we proposed a potential regulatory pathway of TaFH (Fig. 9). The functions of the TaFH gene family were established by multiple endogenous and exogenous factors, and the regulatory network was complex. Therefore, we speculated that these miRNAs of tae-miR167, tae-miR1120-3p, tae-miR9677, tae-miR9679, tae-miR444a, tae-miR172, tae-miR9781, taemiR9675 and tae-miR9661 could inhibit the expression of target TaFHs at the transcriptional or translational levels under low-temperature conditions. This regulation strategy could have contributed to the abnormal cytoskeleton and the meiosis abnormalities, resulting in abnormal pollen. Besides, TaFH genes are associated with the signaling pathway of MeJA, SA, IAA, ABA, and GA. In this study, we found that TaFH1 and TaFH2 were regulated by IAA, TaFH4 and TaFH6 could respond to MeJA, and TaFH3 and TaFH9 were mainly regulated by GA. Collectively, these findings suggested that TaFH1, TaFH2, TaFH3, and TaFH9 might play a crucial role in the regulation of temperatureinduced male sterility in wheat TGMS line BS366.

Conclusions
In this study, we systematically uncovered TaFH genes in the wheat genome. A total of 25 TaFHs were identified, each containing a conserved FH2 domain. The chromosome locations, gene and protein structures, subcellular localization, phylogenetic relationships, Tae-miR167, Tae-miR1120, Tae-miR172, and Tae-miR444 binding sites, and cis-elements were also characterized. The TaFH expression levels in different tissues demonstrate their potential role in flower development. The pollen cytoskeleton of BS366 was abnormal in the three stages under fertility and sterility conditions. Based on quantitative RT-PCR analysis, the tested TaFH genes were highly expressed in inflorescences and response to abiotic stressors. Collectively, these findings establish a foundation for further exploration of TaFH genes and provide novel insights into their biological functions.

Sources of sequence data
The whole-genome sequences of Triticum aestivum L. (T. aestivum, Ta) were retrieved from the wheat genome Ensemble database (http://plant.ensembl.org, IWGSC RefSeqv1.1). To identify members of the formin genes family from T. aestivum, we adopted two methods: Firstly, the hidden Markov model (HMM) profile of the formin (accession number in Pfam: PF06507) was downloaded from the Pfam database (http://pfam.sanger.ac. uk/) [34] and used as a query to search for candidate formin proteins in the T. aestivum proteome using HMMER3.0 [35]. Additionally, Arabidopsis formin protein sequences were adopted as queries to reveal the T. aestivum L. proteome via BLAST. Genes collected from the above two methods were integrated, and duplicates were eliminated manually. All protein sequences of the candidate formin genes were verified using domain analysis programs NCBI-CDD (https://www.ncbi.nlm.nih. gov/cdd/) and SMART (Simple Modular Architecture Research Tool; http://smart.emblheidelberg.de/), the gene structure with intro-exon boundaries of TaFH gene family using the Gene Structure Display Server (GSDS). Basic physical and chemical parameters of TaFH proteins, including amino acid length, theoretical isoelectric point, and molecular weight, were obtained from an online program, ProtParam (https://web.expasy.org/ protparam/).

Homology modeling
The 3D structure of wheat formin proteins was constructed by homology modeling method, which based on the resolved of homologous protein. The Phyre 2 server was been used to obtain the 3D structure and further visualized by PyMOLWin.

Chromosomal distribution
The location of each formin gene on the wheat chromosomes was mapped to IWGSC RefSeq v1.1 (cv. Chinese_ Spring) using blast programs (https://blast.ncbi.nlm.nih. gov/Blast.cgi). The centromere locations were retrieved from IWGSC (www.wheatgenome.org). Then, using the MapInspect tool (http://mapinspect.software.informer. com/), their locations were drawn onto the physical map of each chromosome. The graphic of chromosomal distribution was refined using Adobe illustrator.

Prediction of miRNA target genes
MiRNAs and their targets in TaFH genes were predicted via a web-based psRNATarget server using default parameters [37]. They were then visualized using Cytoscape (version 3.7), as described previously [38].

Identification of the cis-acting elements in the TaFH genes
The cis-acting elements in the 1500 bp upstream promoter regions of the identified TaFH genes were analyzed via PlantCARE (http://bioinformatics.psb.ugent. be/webtools/ plantcare/html/). All of these sequences were used to identify the cis-acting elements by the recently released T. aestivum genome database (http:// plant.ensembl.org/, IWGSC RefSeqv1.1).
For tissue-specific expression analysis, six tissues (root, stem, leaf, glume, stamen, pistil) from BS366 were collected in the field experiment station. Two-week-old wheat seedlings (20°C, 12-h day/12-h night cycle) were used for abiotic stresses analysis, as per the methods described by Bai et al. [28]. In brief, we sprayed BS366 seedlings with 2 mM SA, 100 mM MeJA, 100 mM GA, 50 mM IAA, and 100 mM ABA. For the high-salinity and drought treatments, the roots of wheat seedlings were soaked in 200 mM NaCl and PEG6000 (− 0.5), and leaf tissues were collected from the seedlings 0, 2, 4, 8, 12, and 24 h post-treatment. To evaluate the lowtemperature stress, 2-week-old seedlings were incubated at 10°C, and leaf tissues were collected from the seedlings at 0, 2, 4, 8, 12, and 24 h time points. All samples were frozen in liquid nitrogen immediately after collection. And a flowchart was provided in Fig. S4.

Gene expression assay/analysis
For expression analysis of TaFH genes, total RNA was isolated from wheat tissues using TRIzol reagent (Invitrogen, USA) following the manufactures' instructions. First-strand cDNA synthesis was performed using a PrimeScriptTM RT Reagent kit with gDNA Eraser (TaKaRa, Japan). Next, qRT-RCR assay was conducted on an Eco Real-time PCR system (Illumina, USA) using SYBR® Permix Ex TaqTM (TaKaRa, Japan), as described previously [31].
Relative expression levels of TaFH genes were determined using the comparative threshold cycle method 2 -△△ CT [39]. Wheat 18S gene served as the reference control. All the qRT-PCR reactions were conducted in three biological replicates. The primer used for qRT-PCR were conserved for all three homologues and designed using Primer premier 5.0 program and are listed in Table S2.